An update on Lessirova

[Daphne]

We're still hard at work on Lessirova. It has an entirely new engine under the hood compared to Starter Marriage (let's just say I learned a lot from SM), so that's taking more time than I realized.

And then, there's the art.

Right now, we do not have an artist assigned to it. Art for a multipath game is a hard assignment: you have to be fast, good, and extremely reliable, and there aren't that many artists who hit all three. (Hi, PK!)

We are considering using AI assist for the artwork. We know that this is a hot-button topic, and we'd like to get some feedback about it here.

[jwik01]

When I see an AI-generated image that I recognize as being by AI, it massively cheapens and diminishes the image and whatever it's attached to, in my mind. I don't actually have a philosophical objection to AI images, it just triggers a 'this is garbage' emotional response which makes it very difficult to engage with. They call it slop for a reason: it communicates haste and laziness and lack of effort. That said, if I couldn't tell that something was AI-generated or looked AI-generated, I wouldn't feel cheated if I found out it was. After a certain point it's just another artistic tool. And for all I know I see AI-generated images all the time without realizing it, alongside the ones I do recognize.

[fortemp]

I agree with jwik. I guess if I can’t tell (and you don’t admit it) all’s fair in erotic comics, but part of what I imagine myself paying for here is art made by humans. I can go prompt stable diffusion myself, and for a lot less than a MCC membership…

[BenHur Over]

I've always been somewhat indifferent toward AI-generated images, and based on most of the material I've seen, I don't think the technology is quite there yet.

If a project plans to rely fully on AI images, the quality needs to come close to the consistency and detail in Gossip Goblin's videos, otherwise, I'm not interested. That level of polish is genuinely impressive, but it's clearly difficult to achieve.

For me, the best middle ground is using AI images purely as placeholders for the Multipath Story projects. That way we could get our hands on the next version (or an early build) of SM and Lessirova as soon as possible, while the real artwork is still being drawn and later swapped in by real artists. It would also mean the team wouldn't have to prioritize the games over other comics they're working on, such as SJI.

[Daphne]

If we went the AI route, our target would be something like this: https://www.instagram.com/watch.null.tv/

[sneakybunny]

If we went the AI route, our target would be something like this: https://www.instagram.com/watch.null.tv/

I think same as others have said, I would be happy with AI as placeholder images to ease the pressure on the artists for the multipath stories. Then just replace them when the real artwork is ready.

[jwik01]

If we went the AI route, our target would be something like this: https://www.instagram.com/watch.null.tv/

Those look pretty good! The algorithm seems to work a lot better when it's constrained to a particular style. 99% of the slop I notice as being AI is either 'photo with a lot of artifacting' or 'generic Pixar-esque cartoon,' so if you can avoid that...

[BenHur Over]

If we went the AI route, our target would be something like this: https://www.instagram.com/watch.null.tv/

I think there should be a poll for choosing which AI styles to go with. Looking at the example you provided, they use a lot of different styles, and I do not like most of them at all. Please also avoid the ones used in Rapid Induction, they're awful. I'd much prefer something more realistic, or stylized in a more Western comic way, and less anime inspired.

[arm7man88]

The vast majority of AI created porn, especially in the mc world, is really bad, and complete slop. But there are some I’ve thoroughly enjoyed and these look nice.

[Haight]

I really disagree with LLM-created images (they aren't art) and think the site shouldn't use them.

[USNAWolf]

I personally strongly disagree with any of the use cases for LLMs, particularly for art.

I'll also take the hits for not being a strong fan of the multi-path stories, and even less so as it has seemed that those are pulling the resources away from everything else being created.

[Bob42]

The substantial use of genai images (they are not and can never be art) would mean I would unsubscribe and probably not be back.

[merryweather]

I also really dislike AI art and would not read/play Lessirova if that's what it was.

[ktc1221]

My opinion of AI imagery is that Pandoras box has been opened and not making use of the technology is a fruitless resistance. The artists that were gonna be fucked by it, have been, and so long as people are responsible with it moving forward it greatly diminishes the negative stigma associated with it. That said... I would still prefer no art and inserting human made artwork later over the use of AI at all, even for temporary imagery.

It feels sad that despite years of direct financial support to this site it's still not enough to fund such a project properly.

Would people be willing to have things like... idk just lineart and it gets colored later kinda deal? I understand thats a less finished product but I dont need it to be a full color spread to find it enticing. Take a page out of the Manga industry standards? I suppose thats a poor comparison as they are razor thin profit margins.

Just my two cents

[Daphne]

I've updated the Lessirova Coming Soon page with an example of what the art might look like.